Winning the Linux Wars (MCP)

Linux is a operating system's kernel and it's the one thing deadend-user should never ever be in contact with.

Mac and Windows both are operating systems + operating environments so the comparison is not applicable.

The real fact of life is that majority of people useing computers don't know what they are doing and in the future this will be a fact in any broadly used computer platform.

You haven't used SuSE9/KDE3 lately, have you?

No, it's not identical to Windows, but from experience, I can tell you that it would not be any harder to train... and it would be *much* easier to support, since all users don't automatically have administrative privilege.

Honestly, I'm pretty happy with XP. Or I was until they slipped that "you grant us permission to r00t you whenever we feel like it" clause into the SP2 upgrade EULA.

Don't believe me on that? Ask someone who has to comply with HIPAA.

I think this can and will change, but it will not be an overnight transition.

This is a very big and very common misconception. The UI for every version of MS Windows is far from being user friendly. The only reason that people believe it is is because it's the only UI they've ever seen. From a design standpoint, the Win2k and WinXP UI's are pretty bad.

Folks, when dissing a competing product you play to the customer's fears. This article doesn't mention the relative quality of the user experience at all. So it's pretty safe to assume that this doesn't keep corporate decision makers awake at night.

The article does have some wonderful FUD techniques. (1) Talk to the CEO, not the decision makers. (2) Talk financials last: this allows you to talk up the value of intangibles. (3) Leverage external information which supports your point of view[1]. And so on.

Once you know the major technique outlined in the subtext of the article -- prey on the customer's fears -- then it's easy to use it in reverse.

• Customers are afraid of lack of support -- well tell a tale about a customer with a critical flaw in Windows that was effecting their business and that Microsoft just wouldn't listen (and heh didn't we just have one of those with WMF). And with Linux all is rosy -- if Red Hat won't listen to you, then you can pay a small consultancy that will.

• Access to trained staff? Linux has real certification, based on knowledge. Do you really believe that all those shops displaying "MSCE" banners offer quality training? And you'll need that training, since you'll need to pay new staff to learn to administer Windows -- computer science faculties mainly use Linux now.

• Linux has no secret roadmaps. If a feature is running late you will know. Compare that to the new Window's filesystem that was meant to be in NT, then Xp and then Vista. Microsoft claimed it was "on schedule" and then suddenly it was held over to the next operating system. Do you want your business depending upon that sort of behaviour?

• Linux has no sales force. Microsoft will promise a multiuser MS-DOS if that will get them a sale. You've no idea if Engineering will deliver on Sale's promises (in the case of multiuser MS-DOS v5 they didn't). You can ask Linux developers directly and get a real person giving a real answer.

The other interesting subtext of the article is that the Windows MCPs quoted put Windows first, not the customer who is paying them. There's no need for consultancy firms to have a Windows versus Linux attitude -- it's your dime so they should have a "whatever works best for you" attitude. And so the article forms a list of firms that customers would do well not to employ.

[1] For example, take US CERT's list of software vulnerabilities on its face and say "The US government's prime computer security agency says Windows has ten times less security issues than Linux". The CERT should be slapped for releasing that data without a commentary noting the shortcoming of the raw data for use in analysis.

A good end-user product wouldn't change it's interface constantly, to the point of no recognition by it's users (w2k -> wxp). Oh, and good end-user product would be free of bugs, like the one that won't allow non-administrative user connect with remote server via Citrix Program Neighborhood. And so on...

Then you've never watched a COMPUTER newbie try out both Windows and Linux.

EVERY report I've ever seen says that if you take someone who has never used Windows, and asked them which is easier, Windows or Linux, they go for linux every time. (And IME that's true - when we gave my mum a computer she found linux easier.)

But because Windows is "good enough", when you take someone used to Windows and plonk them in front of linux, they find it hard because it's "different".

Cheers,
Wol

I'm far from a Microsoft fanboy, I just don't like to see people taking biased shots at Microsoft for what I consider a very good end-user product.

I've yet to see a Linux desktop environment that is more user friendly to the common public than the Windows interface (or Mac, also a corporate entity).

It's interesting you should mention the "user-friendly" nature of the MS-Windows UI. First and foremost, the UI on MS-Windows sucks. Consider the start button: how are programs organized? Not by function, but by vendor. This simple little detail belies the major problem with the MS-Windows platform-- it is vendor-oriented, not user-oriented.

That same misdesign philosophy can be seen in applications, from the document format incompatibilities of various MS-Office products to the difficulty in removing Outlook Express from an MS-Windows installation (it's not in the add/remove software control), to deep product branding for all MS-Windows software from any vendor.

The MS-Windows interface is user-friendly to those who already know the MS-Windows interface. My family has been using Linux for years. They despise using MS-Windows machines; as much as I get complaints about both the Mac and Linux interfaces, at least I don't deal regularly with the disgust they express when stuck with MS-Windows. The Mac interface is superior to most Linux desktops; but most Linux desktops beat MS-Windows for ease-of-use and user-friendliness.

Applications may be a different matter. Application consistency within the same platform varies dramatically, so a true study would be difficult, but as another post pointed out, MS-Windows applications are hardly consistent.

The administration of MS-Windows isn't even easier anymore. Once a Linux box is configured, there's nothing more to do, except the occassional update. With MS-Windows, there's the regular updates, as well as anti-spyware and anti-virus protection updates, and the occassional re-install when nothing else seems to fix the system.

Of course, that's just my experience with relatives' machines. Since they are just regular users and don't have all the rules and constraints of a typical business machine, I guess I could just blame them.

1. It would cost over $1 billion (a Gigabuck) to develop this Linux distribution by conventional proprietary means in the U.S. (in year 2000 U.S. dollars).
2. It includes over 30 million physical source lines of code (SLOC).
3. It would have required about 8,000 person-years of development time, as determined using the widely-used basic COCOMO model.
4. Red Hat Linux 7.1 represents over a 60% increase in size, effort, and traditional development costs over Red Hat Linux 6.2 (which was released about one year earlier).

Someone else, inspired by my paper, wrote the very interesting paper Counting Potatoes: The size of Debian 2.2; they found that Debian 2.2 includes more than 55 million physical SLOC, and would have cost nearly $1.9 billion USD using over 14,000 person-years to develop using traditional proprietary techniques.

Obviously Red Hat Linux 7.1 and Debian 2.2 are quite old. Red Hat 7.1, for example, didn't include massive packages like OpenOffice.org. So the actual development costs of current implementations would be even greater. And note that these only estimate replacement cost, not value. A lot of research has gone into developing competing approaches and new ideas, and if only the "winner" went in, the effort that went into the alternatives isn't counted (and it really should be, because those alternatives cost time and caused a better result).

However,
a. It is a good point - WTF do they do with $6B? Compared to the innovation from mere $100's of millions in VC funded startups they appear to be spinning their wheels pretty badly - Losing $100's per unit on a stripped down PC (XBox) hardly qualifies as innovation.

b. The reason that "hardware accelerated" OS chips have not appeared is that "Moore's Law" and DV make them obsolete before they could be finished... I don't know if you have noticed, but the clocks rates stopped going up! The barrier to doing hardware accelerated OSes these days is how to know that they actually work...

The "do it in software" mantra is one of captiulation of the verification gods... It cannot be verified prior to shipping, so we leave room to fix it later...

That is not to say that one does not need "programmability", but neither does one need to be able to program that which has not changed in 25 years...